The vast majority of climate modeling studies treat human behavior as an external, unpredictable factor. They have projected how the climate would change in a variety of possible greenhouse gas emissions pathways, but have not evaluated the likelihood of those pathways.
That approach informs the public and policymakers about what climate paths they should follow in order to achieve the best outcomes for human society and other species, but it does not provide information about which of the nearly infinite possible paths societies most likely will follow. The Intergovernmental Panel on Climate Change (IPCC) uses scenarios ranging from less than 2 degrees Celsius (3.6 degrees Fahrenheit) to more than 4°C (7.2°F) warming above pre-industrial temperatures by 2100, but IPCC does not analyze the likelihood of each outcome.
To address that shortcoming, University of California Davis climate economist Frances Moore led a new study, published in the prestigious journal Nature, that incorporated seven social, political, and technological feedbacks into climate models. It’s an effort to assess which human emissions pathways are the most likely.
This approach “is important for adaptation because increasingly we need to give people information about what climate risks are going to look like over the next 50 or 100 years,” Moore said in a phone interview, which is very difficult if scientists are unable to constrain the likely range of human emissions over that period.
The results of the study provide reason for optimism: The Paris Climate Agreement targets remain within reach in about three-quarters of the 100,000 model simulations run by Moore’s team through the year 2100. While significant uncertainties remain, the study envisions a possible future in which a cascade of social and political and technological feedbacks could lead to an accelerating decline in human greenhouse gas emissions.
The human climate feedbacks
To identify the relevant societal feedback processes that could influence how human greenhouse gas emissions change over the coming decades, Moore and her co-authors conducted a four-day interdisciplinary workshop and a review of literature across relevant fields including social and cognitive psychology, economics, sociology, law, political science, and energy systems engineering. This assessment uncovered seven such climate-social system feedbacks, described below.
The “social conformity feedback” incorporates the influence of public opinion and individual decisions into the model. Public support can often translate into policy changes, with varying success depending on the type of government, and individual behavioral changes can persuade other people to engage in similar actions. For example, if individuals install solar panels or a heat pump or purchase an electric vehicle, those changes can motivate friends or family or neighbors to make similar changes, especially if the individual discusses those decisions with peer groups.
The “climate change perception feedback” accounts for the fact that as climate change and its extreme weather impacts worsen, individuals affected by those events or witnessing the impacts may be more likely to support policies to address the root cause(s). On the other hand, the study authors noted that “several papers have found evidence that interpretations of weather events are filtered through pre-existing partisan identities or ideologies”: As in the analogy of the frog in a boiling pot of water, people may shift their perceptions of what constitutes “normal conditions” as weather gradually becomes more and more extreme over time.
The “political interest feedback,” which describes how policy changes can activate powerful lobbying interests, similarly acts in both positive and negative directions. For example, policies that support renewable energy can bolster the wind and solar industries and their lobbying efforts, but those policies can also trigger adversarial political and public relations activities by powerful fossil fuel interests.
The “credibility-enhancing display feedback” is similar to the individual action component of the social conformity feedback, but it applies to influential individuals. For example, if researchers advocating for climate policies or community ambassadors promoting solar panel installation personally take measures to lower their individual carbon footprints, research has found that the public will view them as more credible, thus enhancing the efficacy of their advocacy.
The “expressive force of law feedback” incorporates into the models that changes in laws and regulations can alter the perception of social norms, attitudes, or behaviors. For example, research has shown that the legalization of gay marriage, smoking bans, and COVID-19 lockdowns had significant effects on public acceptance and norms related to those issues.
The “endogenous cost-reduction feedback” accounts for the fact that as new technologies are increasingly deployed, their costs can potentially fall rapidly as a result of economies of scale, lower input costs, and increased efficiencies that come from the economic theory of “learning by doing.” This effect has been demonstrated, for instance, by the plummeting costs of clean technologies such as wind turbines, solar panels, and lithium batteries.
Finally, the “temperature-emissions feedback” addresses direct impacts climate change will have on the economy. Several studies have suggested that worsening climate change damages may slow economic growth, and given that economic productivity is connected to energy use, this effect could also slow greenhouse gas emissions growth. On the other hand, rising temperatures will increase energy demand for cooling. One 2019 study estimated that these effects combined would reduce greenhouse gas emissions approximately 3% per degree Celsius of warming.
Results inspire hope, justify cautious optimism
The study authors performed 100,000 model simulations incorporating these seven climate-social feedbacks and then clustered together model runs with similar trajectories of climate policy and emissions through the end of the century. They found that the most common cluster (called the “modal path”), representing nearly half of the model runs, resulted in a most likely warming of about 2.3°C (4.1°F) above pre-industrial temperatures in the year 2100. The second-most common group of simulations, representing more than a quarter of model runs, was categorized as an “aggressive action” scenario in which governments are successful in meeting the Paris Climate Agreement target of limiting global warming to less than 2°C (3.6°F).
They labeled the third-largest cluster “technical challenges,” representing almost one-fifth of model runs. In these scenarios, government climate policies are similar to those in the most common “modal” cluster resulting in 2.3°C warming, but clean technologies remain relatively expensive in light of a weak “learning by doing” feedback relating to the economies of scale, which would slow efforts to reduce greenhouse gas emissions. In this scenario, global temperatures rise by about 3°C (5.4°C) above pre-industrial temperatures in 2100.
Together, these three scenarios account for nearly 95% of the model simulations, and they all envision many governments enacting climate policies well beyond their current status quo in an effort to meet Paris commitments. Two other clusters labeled by the study authors as “delayed recognition” and “little and late” involve less aggressive government action, as the names imply, but were represented by only about 5% of the model simulations, most likely representing about 3-3.5°C (5.4-6.3°F) warming in 2100.
The study did not include the potential for carbon dioxide removal (CDR) from the atmosphere. Incorporating successful CDR on the scale recommended by the National Academy of Sciences and IPCC (approximately 750 billion tons of cumulative carbon dioxide removed by 2100) would roughly bring the warming experienced in the modal path in line with the Paris Climate Agreement target of limiting global warming to less than 2°C (3.6°F). Combined with the “aggressive action” cluster, this leaves about three-quarters of model runs within reach of the Paris targets.
Key factors and caveats
The most influential factors in the model runs were the strength of public opinion, clean technology cost reductions, responsiveness of political institutions, and the role of cognitive biases. Several of these factors tend to act against climate solutions in the United States, with its population heavily politically polarized and government policy not very responsive to public opinion in any event, perhaps due largely to structural status quo biases. But rapidly falling costs have led to clean technology adoption even in many so-called “red” states, and a number of other states have carried out ambitious local climate policies. And in many countries like Canada and across Western Europe, governments have been responsive to public support for climate solutions.
As in any modeling exercise, there are significant uncertainties in the study’s results, related both to physical and social factors. It’s also possible that the study did not include some important negative feedbacks, for example if rising energy prices reverse public opinion support for climate policy, or if highly politicized nationalism or isolationism increases as climate change potentially makes resources increasingly scarce.
Climate modeler Drew Shindell, not involved in the study, also flagged potential unaccounted-for negative feedbacks like media biases, noting, “for example there are loads of reports on EV battery fires (even though they are far less common than gas-powered vehicle fires), which could lead to public opinion being a negative feedback suppressing changes.”
Consilience with cautious optimism from other studies
The study’s bottom line could best be characterized as suggesting that governments might succeed in implementing climate policies starting within the next decade that are sufficient to limit global warming to somewhere between a bit less than 2°C and 3°C by 2100. While this includes a wide range of increments of climate change and associated damages, the results suggest that the most catastrophic outcomes above 3°C are relatively unlikely, with the aforementioned caveats.
As climate scientist and past regular contributor to this site Zeke Hausfather found in a separate Nature commentary published with Moore, numerous recent studies have found that when accounting for implemented climate policies and pledged targets, the most likely temperature outcomes tend to fall in that same 2-3°C range. As Hausfather noted via email, Moore’s new study is “a completely independent way of assessing plausible emissions outcomes … which makes it valuable and its consilience with other lines of evidence noteworthy.”
The new study’s most likely outcome suggests that global emissions will continue to rise over the next 8 years, missing the 2030 Paris pledges, but will decline rapidly thereafter to bring the 2050 targets within reach. For example, increasing public support
- could spur climate policies in many countries,
- which could catalyze the deployment of clean technologies,
- which could drive down their costs,
- which could further increase public support,
- … and so on in a cascading positive feedback effect.
As Moore told Yale Climate Connections, near-term climate failures could make people “way too pessimistic. If there are these spillovers like ‘learning by doing’ effects and cost reductions and network effects, it’s possible to be taken by surprise positively” by accelerating emissions reductions after 2030.
Given ongoing struggles and delays in adopting and implementing climate policies in many countries like the United States, some advocates for climate action may be heartened by this potential hope for success – or at least for progress – in coming decades.